Schema.fiware.org: FIWARE Harmonized Data Models presentation, by Jose Manuel Cantera Fonseca. How-to sessions. 1st FIWARE Summit, Málaga, Dec. 13-15, 2016.

1. schema.fiware.org
José Manuel Cantera Fonseca
Technological Expert. FIWARE Data Models Technical Lead.
josemanuel.canterafonseca@telefonica.com

2. Background
Introduction to schema.org
2

3. Introduction
• What is schema.org? http://schema.org
o A collaborative, vendor-neutral, community-based initiative targeted to provide a
shared set of harmonized, cross-domain vocabularies for publishing and consuming
structured data on the web
o Enables structured data exchange in an interoperable fashion
o In widespread use today (>12 million sites using it and growing everyday)
• Why schema.org?
o search engines can understand “the meaning” of web content
o richer and more interactive user experiences when searching for information
o interoperability between the Web and/or other applications
calendars, email clients, personal assistants, …
● Who is behind schema.org?
○ Major search engine providers (Google, Yahoo!, Microsoft, Yandex, …) which act
as project sponsors
○ A community of developers / domain experts making contributions
○ W3C Community Group https://www.w3.org/community/schemaorg/
■ A CG is not governed by the official W3C Process

4. Introduction (II)
Nowadays: Web search has
turned out to be the best
consuming application for the
graph of knowledge

5. Introduction (III)
• In order to provide any kind of structured data, it is needed:
o Data Model (structural organization of data):
Graph vs DLG vs Tree vs Entity-Attribute vs vertical-specific ...
o Syntax (underlying format used to represent such data structure):
JSON vs XML vs RDF vs HTML5 vs JSON-LD vs RDFa vs Turtle …
o Vocabulary (dictionary):
Terms → Entities (types, classes), relations (attributes, properties) and
enumerations (tabulated values)
o Identifiers for instances (entity ids)
● Many years of formal standards work, battles, bells & whistles …
○ The stick

6. schema.org - Facts
• Work started in August 2010
o Sponsored Microsoft, Google, Yahoo!, Yandex
• Goals:
o Make it very easy for the webmaster
write once for all search engines using one vocabulary (single schema)
o Make it possible to use it in combination with other vocabularies
o Wide range of cross-domain topics (297 classes, 187 relations, initially)
people, places, events, products, offers, recipes, businesses, opening hours ...
• Successful initiative, de facto standard, thanks to:
o Being the substrate behind contemporary web search engine experiences
o Supported by popular CMS tools and multiple verticals
Drupal, Wordpress, Blogger, Youtube
Ticketmaster, ebay, alibaba, BBC, European Broadcasting Union, etc.
o Flexibility, pragmatism, agility, maintainability,

7. schema.org - Structure and core data model
• schema.org core data model is a graph, roughly, similar to RDF
• schema.org defines the following term categories:
o Data types (Number, DateTime, Text, URL, etc.)
o Entity types (Event, Restaurant, Person, Offer, Place, etc.)
Arranged in a multiple inheritance hierarchy
Currently 583 (core vocabulary)
o Relations (properties)
multiple domain and range
Currently 846 (core vocabulary)
o Enumerations
wherever there are a limited number of typical values for a property, there is a
corresponding enumeration specified in schema.org
Ex. http://schema.org/ItemAvailability
Currently 114 (core vocabulary)
• schema.org is not intended to be a universal ontology.
o it should be used in combination with other vocabularies

8. schema.org - Design principles
• Avoid the over generalization trap
o Concrete classes and properties. Escape from terms like ‘Agent’.
o Try to find the right initial level of generality
• Avoid the mess of unique URIs for each instance
o Use only unique URIs for schema.org terms
o Reconcile entities as per attribute values
• Incremental complexity. Start simple. Go back and fill in the blanks
o And let applications ask for sophistication later, after mass adoption
Adding new subtypes or more descriptive properties
o Not trying to find the perfect model
respond to the demands of applications and data providers
o Move fast. Iterate fast. Accept mistakes
Publish schemas early even though it is known that improvements will be needed
• Pragmatic Extensibility
o A small core vocabulary for each topic and rely on extensions for fine grain details
o Work together with interest communities

9. schema.org - Application example (I)
JSON-LD representation
(Web page)
Search page enriched snippets

10. schema.org - Application example (II)
enriched Gmail inbox
JSON-LDrepresentation(emailcontent)

11. FIWARE + schema.org
11

12. FIWARE + schema.org - Introduction (I)
• Idea :
o apply all the successful design principles and workflows of schema.org to
produce harmonized schemas in the "Internet of Things" (IoT) area
enabling semantic interoperability that bridges IoT and non-IoT applications
real scope is “real world things” rather than Internet of Things.
• In cooperation with GSMA and mobile operators
o http://www.gsma.com/connectedliving/wp-content/uploads/2016/12/CLP.26-v1.0.pdf
• IoT + schema.org complementary initiative introduced by Dan Brickley
(schema.org chairman) through the following position paper:
o https://docs.google.com/document/d/1mE2hN83lXzqXA-
RQYMURJxIlEktcZkFB252BQCCnkKk/edit#

13. FIWARE + schema.org - introduction (II)
• Does this really make sense?:
o (IoT) data provider role, publishing data using iot.schema.org, will be similar
than the role played by a Webmaster
o Application / (IoT) data consumer, role, will be similar than the search engine,
email client,etc. roles.
Dashboards, smart (interoperable) applications, etc.
• We can bring all the schema.org benefits from the Web to the IoT-Big
Data domain
o avoiding silos, fragmentation, etc.
o Give carrots to data / IoT providers instead of sticks (ex. lengthy ontologies)
o “We give value to your data if it is harmonized”
o IoT and non-IoT entities can be easily “linked” following schema.org principle
of “local view” but a “single”, coherent vocabulary.

14. 14

15. Entity modelling with schema.fiware.org

16. schema.fiware.org .- An open source project
Contributions
are welcome!

17. Questions
17

18. References
• http://schema.org/docs/documents.html
• https://www.w3.org/community/schemaorg/how-we-work/work-in-progress-mechanisms-
webschemas-and-the-pending-area/
• https://www.w3.org/community/schemaorg/how-we-work/issue-management-on-github/
• http://events.linkeddata.org/ldow2014/slides/ldow2014_keynote_guha_schema_org.pdf
• http://queue.acm.org/detail.cfm?id=2857276
• https://search.google.com/structured-data/testing-tool
• https://developers.google.com/gmail/markup/reference/
• https://msdn.microsoft.com/en-us/library/dn632191.aspx
• https://developers.google.com/search/docs/guides/enhance-site
• http://schema.fiware.org
• http://webschemas.org/docs/cg/sdo-wot-tpac-2016-09.pdf
• https://docs.google.com/document/d/1mE2hN83lXzqXA-
RQYMURJxIlEktcZkFB252BQCCnkKk/edit#
• https://developers.google.com/knowledge-graph/

19. Thank you!
http://fiware.org
Follow @FIWARE on Twitter